Update README

This commit is contained in:
Gregory Nutt 2015-05-24 17:07:34 -06:00
parent c5fc9af3a5
commit 36d8df161b
1 changed files with 40 additions and 34 deletions

View File

@ -16,6 +16,7 @@ The SAMD20 Xplained Pro Starter Kit may be bundled with three modules:
Contents
^^^^^^^^
- STATUS
- Modules
- Development Environment
- GNU Toolchain Options
@ -27,6 +28,45 @@ Contents
- SAMD20 Xplained Pro-specific Configuration Options
- Configurations
STATUS/ISSUES:
1. The FLASH waitstates is set to 2 (see include/board.h). According to
the data sheet, it should work at 1 but I sometimes see crashes when
the waitstates are set to one (about half of the time) (2014-2-18).
2. Garbage appears on the display sometimes after a reset (maybe 20% of
the time) or after a power cycle (less after a power cycle). I don't
understand the cause of of this but most of this can be eliminated by
simply holding the the reset button longer and releasing it cleanly
(then it fails maybe 5-10% of the time, maybe because of button
chatter?) (2014-2-18).
- The garbage is not random: It is always the same.
- This is not effected by BAUD rate. Curiously, the same garbage
appears at different BAUD settings implying that this may not even
be clock related???
- The program seems to be running normally, just producing bad output.
3. SPI current hangs so not much progress has been made testing the I/O1
module. The hang occurs because the SPI is waiting for SYNCBUSY to
be cleared after enabling the SPI. This even does not happen and so
causes the hang.
Another note: Enabling the SPI on SERCOM0 also seems to interfere
with the USART output on SERCOM4. Both symptoms imply some clock-
related issue.
The configuration suggests CONFIG_MMCSD_HAVECARDDETECT=y, but as of
this writing, there is no support for EIC pin interrupts.
4. OLED1 module is untested. These instructions were just lifted from
the SAM4L Xplained Pro README.txt file.
5. As of 20154-05-24, this SAMD20 configurations are broken. This
seems to have occurred as a consequence of the SAML21 integration
between NuttX-7.9 and NuttX-7.10. Hence, I expect that the problem
is trivial but nevertheless real.
Modules
^^^^^^^
The SAMD20 Xplained Pro Starter Kit is bundled with four modules:
@ -857,37 +897,3 @@ Configuration sub-directories
This is clearly some issue with initializing, un-initializing, and
then re-initializing. If you want to fix this, patches are quite
welcome.
STATUS/ISSUES:
1. The FLASH waitstates is set to 2 (see include/board.h). According to
the data sheet, it should work at 1 but I sometimes see crashes when
the waitstates are set to one (about half of the time) (2014-2-18).
2. Garbage appears on the display sometimes after a reset (maybe 20% of
the time) or after a power cycle (less after a power cycle). I don't
understand the cause of of this but most of this can be eliminated by
simply holding the the reset button longer and releasing it cleanly
(then it fails maybe 5-10% of the time, maybe because of button
chatter?) (2014-2-18).
- The garbage is not random: It is always the same.
- This is not effected by BAUD rate. Curiously, the same garbage
appears at different BAUD settings implying that this may not even
be clock related???
- The program seems to be running normally, just producing bad output.
3. SPI current hangs so not much progress has been made testing the I/O1
module. The hang occurs because the SPI is waiting for SYNCBUSY to
be cleared after enabling the SPI. This even does not happen and so
causes the hang.
Another note: Enabling the SPI on SERCOM0 also seems to interfere
with the USART output on SERCOM4. Both symptoms imply some clock-
related issue.
The configuration suggests CONFIG_MMCSD_HAVECARDDETECT=y, but as of
this writing, there is no support for EIC pin interrupts.
4. OLED1 module is untested. These instructions were just lifted from
the SAM4L Xplained Pro README.txt file.